Sign inSign up
ModelPilotAI
Product Demo May 2026 v0.9.0
AI Chatbot Infrastructure
Ship AI chatbots
in minutes,
not months.
Connect any model · Upload knowledge · Deploy anywhere
Track everything — no code required.
500+
Teams building
2.4M
Messages handled
8
AI providers
$0
To get started
<5m
Time to first bot
The Problem
Building AI chatbots is
still painful in 2026.
3mo
Time to deployment
Average time to get a custom AI chatbot live — gluing together APIs, infra, and front-end from scratch.
$180K avg engineering cost
7+
Tools duct-taped together
Vector DB · LLM provider · embedding model · widget · analytics · agent framework · webhook system — all separate.
No unified platform exists
2.4x
Cost overrun rate
Token costs, infra bills, and API fees spiral when teams build custom — without usage tracking or budget caps built in.
Zero observability out of the box
The Solution
One platform.
Any model. Five minutes to deploy.
ModelPilot replaces the entire chatbot stack — from knowledge ingestion to deployment to analytics — in a single product.
Knowledge
PDF · URL · FAQ
→
RAG Index
Qdrant · embeddings
→
LLM Router
LiteLLM · 8 providers
→
AI Agents
FAQ · Support · Sales
→
Widget
<4KB · any site
→
Analytics
Cost · sentiment · logs
The full product suite
6 integrated modules
01
Chatbot Builder
Multi-model bots with system prompts, temperature, fallback, and human handoff. Live preview powered by real AI.
02
Knowledge Base
Upload PDFs, scrape URLs, write FAQs. Auto-chunked, embedded, and retrieved via semantic search.
03
AI Agents
FAQ, Support, Sales, Onboarding, Language, and Handoff agents — each with configurable confidence thresholds.
04
Analytics
Message heatmap, cost tracking, model distribution, sentiment scoring, and top conversations by token usage.
05
Integrations
Webhooks on any event. One-click connect to n8n, Slack, Zapier, Make, HubSpot, and Zendesk.
06
Embed Widget
Vanilla JS, <4KB gzipped, zero dependencies. One script tag. Customisable accent, position, and greeting.
Chatbot Builder
Build any bot.
Any model.
✓
Multi-model support
GPT-4o, Claude 3.5, Gemini, Groq, Ollama + more
✓
Real AI test preview
Chat with your bot live using real AI — before deploying
✓
Human handoff + web search
Configurable feature toggles per bot
✓
Persona & tone editor
Name, avatar, greeting, suggested prompts
Configuration
Personality
Test Preview
Support Bot
Configuration
Personality
Test Preview
Status
Live
Bot Name
Support Bot
AI Model
GPT-4o
System Prompt
You are a helpful support assistant. Be concise, accurate, and friendly...
Temperature · 0.7
Fallback
Escalate to agent
Language
Auto-detect
Features
Human Handoff
Web Search
Knowledge Base
Semantic retrieval at
every conversation.
335 chunks indexed
Average score: 0.82 similarity
📄
Source
PDF · URL
FAQ · DOCX
→
Chunker
256-token
segments
→
Embedder
text-embed
-3-small
→
Qdrant
Vector store
per bot
→
Retrieval
Top-5 · score
threshold 0.75
→
LLM Reply
Context-aware
streaming
AI Agents
Six agent types.
Zero code.
Powered by LiteLLM + Flowise
FAQ Agent
Strict RAG — answers only from your knowledge base. Returns "I don't know" when context is missing.
RAG Strict
Support Agent
Handles issues end-to-end. Escalates to human when confidence drops below threshold.
Multi-turn Handoff
Sales Agent
Qualifies leads, answers pricing questions, books demo calls via Calendly integration.
Lead capture CRM sync
Language Agent
Auto-detects user language and responds in kind. Supports 50+ languages and RTL scripts.
Auto-detect 50+ langs
Onboarding Agent
Guides new users through product setup with interactive checklists and progress tracking.
Checklist Step-by-step
Handoff Agent
Detects frustration signals via sentiment analysis and transfers to a live agent with full context.
Sentiment Context pass
Analytics
Every conversation.
Measured.
Resolution
91%
↑ 2% week over week
Cost / conv.
$0.04
↓ 3% vs last month
Heatmap · Bar charts · Sentiment
Hour-of-day activity heatmap, per-model cost distribution, real-time conversation feed
Messages
8,042
↑ 12%
API Cost
$41.90
↑ 8%
Active Users
318
↑ 8%
Message Volume — Last 7 days
Daily breakdown
Mon
Tue
Wed
Thu
Fri
Sat
Sun
Model Distribution
GPT-4o 44%
Claude 28%
Gemini 18%
Other 10%
Infrastructure
Open-source stack.
Production-grade.
100% open-source core
Self-hostable · no vendor lock-in
FastAPI
Backend API · Python 3.12
LiteLLM
AI gateway · 8 providers
Qdrant
Vector DB · semantic search
Next.js 14
Frontend · App Router
Supabase
Auth · PostgreSQL · RLS
Redis
Cache · rate limiting
Flowise
Agent orchestration
n8n
Workflow automation
Railway / Vercel
Deploy · CI/CD
Pricing
Start free.
Scale as you grow.
No hidden fees · No per-token markup
Starter
$0
/month
✓
3 Chatbots
✓
5,000 messages/mo
✓
1 AI provider
✓
Embed widget
Free forever
MOST POPULAR
Pro
$79
/month
✓
10 Chatbots
✓
50,000 messages/mo
✓
All AI providers
✓
Analytics + Webhooks + API
✓
Human handoff
Upgrade to Pro
Enterprise
Custom
✓
Unlimited bots
✓
SSO / SAML 2.0
✓
Audit logs + SLA
✓
Dedicated CSM
Contact sales
Traction
Teams building on ModelPilot AI
500+
Across SaaS, e-commerce, healthcare, and education
Messages handled
2.4M
Across all bots on the platform
Platform uptime
99.9%
LiteLLM gateway SLA
Time to first bot
<5m
Median onboarding time from sign-up
Roadmap
Six weeks to production.
Solo founder · ~40h/week
WK 1
Foundation
Next.js + FastAPI · Supabase auth · LiteLLM gateway · JWT middleware
✓ Done
WK 2
Knowledge Base
Qdrant setup · PDF ingestion · URL scraper · embedding pipeline · RAG retrieval
✓ Done
WK 3
Chat Widget
Vanilla JS widget · EventSource streaming · CORS hardening · session persistence
In progress
WK 4
Agent System
FAQ + Support agents · sentiment scoring · Flowise · n8n webhooks
Planned
WK 5
Analytics & Billing
Token tracking · dashboard charts · Stripe integration · usage limits
Planned
WK 6
Launch Prep
Rate limiting · Sentry · self-serve onboarding · Railway + Vercel deploy · beta
Planned
ModelPilot AI
Start building
today.
No credit card. No infra setup. Your first bot is five minutes away.
modelpilot.ai
Get started free
View live demo
github.com/modelpilot
docs.modelpilot.ai
Confidential · May 2026